DYNAMIC YARD TEMPLATE DESIGN TO IMPROVE UTILIZATION …

JOURNAL OF ENGINEERING AND MANAGEMENT IN

INDUSTRIAL SYSTEM VOL. 6 NO. 2 YEAR 2018

e-ISSN 2477-6025

DOI:10.21776/ub.jemis.2018.006.02.2

Cite this Article as Nurminarsih, S., Putri, M., & Rusdiansyah, A. (2018). DYNAMIC YARD TEMPLATE

DESIGN TO IMPROVE UTILIZATION AND OPERATION EFFICIENCY ON VERTICAL LAYOUT-CONTAINER YARDS CONSIDERING UNCERTAINTY OF TRUCK ARRIVAL. Journal of Engineering

and Management in Industrial System, 6(2), p74-85.

Paper Accepted : Nov, 17th 2018

Paper Published : Dec, 31st 2018 74

DYNAMIC YARD TEMPLATE DESIGN TO IMPROVE

UTILIZATION AND OPERATION EFFICIENCY ON VERTICAL

LAYOUT-CONTAINER YARDS CONSIDERING UNCERTAINTY OF

TRUCK ARRIVAL

Siti Nurminarsih1)

, Maulin Masyito Putri1)

, Ahmad Rusdiansyah2)

Department of Logistics Engineering, Universitas Internasional Semen Indonesia1)

Department of Industrial Engineering, Institut Teknologi Sepuluh Nopember2)

Abstract Loading containers into a vessel can be delayed because of inefficient processes in the container

yard. These container yard inefficient processes can be caused by inappropriate stacking position and

inefficient crane operation. Inappropriate containers stacking positions may need some preparatory jobs of

rearranging containers ahead of loading time to reduce delays in the process. The method will allow

preparatory jobs to be scheduled together with main jobs to optimize use of the idle time of Automated Stacking

Cranes (ASC). This research aims to find a better scheduling strategy of twin ASCs as well as improving land

utilization in a vertical-layout container yard to minimize delays in containers loading process. The uncertainty

of truck arrival is accomodated by implementing iterative rescheduling in look-ahead planning horizon. The algorithm finds the optimal position of the container in the yard by implementing scoring function for each

available slot. Finally, some numerical experiments are conducted to prove the performance of the algorithm.

Keywords: Vertical Layout-Container Yard; Land Utilization; Iterative Rescheduling

1. Introduction International trade is growing during the

economic globalization era. The growth of

international trade brings along an increase in utilization of maritime logistics as the main

means of transportation to distribute goods

across countries. Being passed by 40 percent of

maritime global trade route, making Indonesia as an important part of international trade. That

means Indonesia should prepare their logistics

infrastructures to be efficient and reliable for serving the operations. Unfortunately,

Indonesian performance on logistics is still

lacking and being ranked as 46th out of 160 countries surveyed by World Bank in 2018 [1].

The logistics performance score of Indonesia,

which is 3.15, is lower than other neighborhood

countries such as Singapore (4.00), Thailand (3.41), Vietnam (3.27), and Malaysia (3.22).

This will be disadvantage for Indonesia to

compete globally in economics.

* Corresponding author. Email : [email protected]

Published online at http://Jemis.ub.ac.id

Copyright ©2018 JTI UB Publishing. All Rights Reserved

Based on this issue and the challenge of

global economics growth, the government of Indonesia set up the Masterplan for The

Acceleration and Expansion of Indonesia’s

Economic Development (MP3EI) 2011-2025. One of the important elements of this program is

to enhance connectivity between regions in

Indonesia and internationally through maritime

transportation systems. Each region in Indonesia has different part in economic, based on their

potentials, and connected by maritime

transportation routes called “Sea Toll Road” [2]. These routes connect regions in Indonesia using

hub and feeder concept. By this program, the

transportation access to all regions in Indonesia will be available regularly on definite schedules.

The impact of international trade growth and

Sea Toll Road program increases the container traffic

among ports in Indonesia. Ports should improve their infrastructure and operational efficiency to deal with

this increasing traffic. In fact, the average dwelling

time of ports in Indonesia is above 3 days [3]. Dwelling time is defined as the average time of

container being in the port. The causes of this long

dwell time are the complicated bureaucracies, the scarcity of land for container yard and a container

transfer inefficiency during loading process [4].

http://jemis.ub.ac.id/


INDUSTRIAL SYSTEM VOL. 6 NO.2X YEAR 2018

e-ISSN 2477-6025

DOI:10.21776/ub.jemis.2018.006.02.2

75

Operational constraints at the container yard are

usually caused by a scarcity of the land and

infrastructure in the port. An efficient yard management is needed in the condition of increasing

demand and the scarcity of the land and

infrastructure. To implement an efficient yard

management, it is important to generate yard

template which can respond to uncertainties of operational condition. Yard template is a

guideline of yard operator to determine the

position of incoming container on the yard. The

yard template of vertical-layout container yard should be dynamic because the position of

containers can change all the time while in

container yard. This research developed an algorithm for determining dynamic-yard

template while considering the scheduling of

ASCs and optimizing space utilization. The rest of this paper is organized as follows. Section 2

provides the literature review. Section 3 details

the methods of the problem. Section 4 explains

the numerical experiments and results. Finally, section 5 gives conclusion and possible future

development of the research.

2. Literature Review

Several studies have been conducted to find

better method and approach of yard management.

Most of the research focus on the factors that cause inefficiency in the yard area, including the number of

containers reshuffle and traffic congestion. Yard

planner put containers on the yard based on position spesified on the yard template. The quality of yard

template planning affects the efficiency and

productivity of container handling. Generally, a yard template is made in the planning phase when the ship

had not arrived yet. Whilst when it comes to the

operating phase, some adjustments are often made to

the original layout because of high uncertainties, such as the arrival of the ship, the arrival of the truck, and

the volume to be transported. To improve the

efficiency, planning at the container yard should consider the uncertainties that occurs during

operation level [5].

To improve efficiency in the yard area, Lee et al. [6] combined the optimization of the reshuffle and

traffic congestion to make yard planning of

transshipment containers. A consignment strategy has

been introduced to minimize reshuffle by locating export and transshipment containers at the certain

area dedicated for the ships. The probablity of traffic

congestion has been minimized by implementing “High Low Workload Traffic Balancing Protocol”. A

mixed integer linear programming (MILP) then

developed to determine the crane scheduling. The

implementation of dedicated-yard area increases container accessability, but on the other hand also

decreases yard utilization because of empty spaces

that arise due to the absence of containers. In 2012, Jiang et al. [7] continued the research by suggesting

the space-sharing concept as the solution to the lack

of yard utilization that caused by the implementation of consignment. Jiang et al. [7] created a framework

for allocating space-sharing plan for each ship arrival.

However, this research worked at the planning level

where uncertainties in the operational level are ignored.

Uncertainties are often occured at the

operational phase, for example due to delays in operations at the previous port, weather, arrival of

external trucks, and number of receiving and

delivering containers. The longer the vessel’s path, the higher the level of uncertainty. A few studies have

examined port operations under uncertain

environments [8]. Most of them studied the impact of

vessel arrival time and operation time in quay or berth areas. Han et.al [9] and Zhen et.al [10] studied about

berth allocation and quay crane scheduling problem

under those uncertainties. At the yard areas, Zhen [5] addressed on the problem of planning a robust yard

template considering uncertainty in number of

receiving and delivering containers. These

uncertainties affect the determination of sub blocks in container yard (CY) which are assigned in fully-

dedicated or sharing mode for each vessel. In this

research, vessels are scheduled to visit the port periodically (weekly, 10-days, or biweekly). The

number of receiving and delivering containers from

(or to) a vessel also fluctuates in each period. The traffic congestions in the yard and multiple schedule

cycle time for vessel arrival pattern are also

considered in the developed model. Both of Jiang et

al. [7] and Zhen [5] developed method to optimize the utilization of horizontal layout-container yard which

is commonly used in Asian ports. The operational of

vertical layout-container yard is different with horizontal layout. It is hard to determine the dedicated

area on the yard for certain vessel because the

containers are moved several times at the yard since arrival until departure.

The movement of the container on the vertical

layout-container yard can be divided into three types

of movements. The first one is movement that occurs when container first arrives at the yard, both carried

by external truck or AGV. Next movement is

rehandling. This movement is needed when the container to be retrieved next is not at the top of a



e-ISSN 2477-6025

DOI:10.21776/ub.jemis.2018.006.02.2

76

stack which causes relocation of all containers above

the target container. Rehandling is the major source of

loading delays [11]. Another factor that causes delay is the traveling distance of ASC to move container to

be loaded. To minimize loading delays due to these

two factors, the third type of movement arises, namely remarshaling. Remarshaling rearranges the

containers in some appropiate ways well ahead of

their loading time [11]. Park et al. [12][13] suggested the remarshaling jobs that contribute the most to the

reduction of future loading delays to be selected for

the mix of job queue. Therefore, starting and stoping

time of remarshaling jobs should be determined manually and not yet scheduled together with the

current requested jobs. Later in 2013, Choe et al. [11]

proposed method to schedule the twin ASCs in such a way that the remarshaling jobs are performed

intermittently while the ordinary requested jobs are

performed. Adopting the predetermined fixed-length look-ahead horizon in a regular interval framework

by Park et al. [14], the remarshaling jobs are mixed to

the requested jobs unless the load of requested jobs is

heavy. The objective function of this method is to minimize the total waiting time of external truck,

delay of AGV, and total makespan of the schedule.

The utilization of space is not a major consideration in this research.

Previously, Nurminarsih et al. [4] has been

developed a sharing-area framework for vertical

layout-container yard under uncertainty of the arrival of containers which is adopting the method

developed by Jiang et al. [7] and Zhen [5]. The

research is developed based on real situation in one of port in East Java, Indonesia, which implement

vertical layout-container yard. The uncertainty of

container arrival is translated into container arrival probability during the slot opening horizon, 5-days

ahead the vessel arrival schedule. This probability is

based on historical data of container arrival.

Nurminarsih et al. [4] only introduced the framework and method for implementing the space-sharing

concept without considering the movement of ASC.

This research aims to develop the previous research of Nurminarsih et al. [4] by taking into account the

scheduling of ASCs. Model for scheduling ASCs is

adopted from Choe et al. [11] while adding the consideration of space utilization as the outcome of

sharing-area concept.

3. Methodology

3.1 Operation in Vertical Layout-Container

Yard

Vertical layout-container yard is often

called the European Layout, because this layout is generally used in European container ports. In

the vertical layout-container yard, the yard has

blocks positioned perpendicular to the quay (see Fig. 1) and the I/O points are located at both ends

of the blocks to handle storages and requests

from the seaside and landside [15]. Based on the the direction of flow, the containers handled on

the yard terminal can be categorized into three

groups: export, import, and transhipment

containers. An export container carried in by external truck (ET) and being stored in the yard

until the arrival of carrier ship, the container then

being moved from the yard and loaded into the ship by Quay Crane (QC). On the contrary, the

import container carried in to the yard by

Automated Guided Vehicle (AGV) after being unloaded from the ship by QC. The import

container will be stored at the yard until carried

by ET. A transhipment container is also unloaded

from ship and carried in by AGV to the yard, but the container will be carried out again by AGV

to be reloaded to another ship instead of being

carried out by an ET. The movements of a container in the yard

are handled by Automated Stacking Crane

(ASC). There are two identical ASCs for each

block in the yard, namely seaside and landside ASC. Seaside ASC works on the quay side,

assigned to deliver (or take) a container to (from)

AGV. On the other hand, the landside ASC is assigned to carry in (out) a container to (or from)

yard from (or to) an ET on the landside H/P.

Because the size of both ASCs are identical, the ASC cannot move through each other and

separate by a transfer area which is called

handshake area. Unlike the transhipment

container, both import and export container enter a block through one end of a block and go out

through the opposite end of the block, this means

that both import and export container cannot help being handled by both seaside and landside

ASC. The job of moving container from landside

H/P to the yard or vice versa is called lanside job. While the job of moving container from seaside

H/P to the yard or vice versa is called seaside job.

Both landside and seaside job are categorized as

main job to differentiate it from preparatory job: rehandling



e-ISSN 2477-6025

DOI:10.21776/ub.jemis.2018.006.02.2

77

Fig 1. Container Yard Configuration

and remarshaling. The main job is given top priority to be done by an ASC, unless some

preparatory jobs are needed, for example the

targeted container is not on the top of the stack so some rehandling jobs are required to be done

first. The efficiency of loading operation is

considered the most critical to the productivity

of the whole terminal because the berthing time of a ship depends on how fast the containers are

loaded [11].

In this research, we observe one of international container port in Indonesia which

has yard area consists of 6 blocks: 3 blocks for

domestic shipping and 3 blocks for international shipping. Each block consists of 9 containers

width (row), 39 containers length (bay), and 5

container stacking height (tier). One slot

container can accommodate one 20-feet container. Figure 1 shows the configuration of

container yard in this port.

This port implements consignment strategy, which the agent books some slots for upcoming

vessel 6 days prior to vessel departure. The booked space in the yard will be assigned to a

vessel for 5 days before the vessel departure.

From the provided data, we obtain that higher workload of incoming containers always happen

near vessel departure date, even some containers

arrive when ship is already docked. The

distribution of the container arrival follows pattern as shown in Figure 2. This behavior leads

consignment strategy produce high

inefficiencies in the operation because of the unoccupied booked spaces. In fact, the real

arrival containers sometimes different from the

booked containers. Because of this problem, yard operators usually open less than 100% of

the booked spaced on early booking day. At first,

30% of the booked spaced are opened and the

percentage of opened booked space will be added if the incoming containers increase [4]. By

this strategy, the land efficiency will increase,

but some containers for the same vessel may be placed far from others. This will



e-ISSN 2477-6025

DOI:10.21776/ub.jemis.2018.006.02.2

78

Fig 2. The Pattern of Containers Arrival [7]

cause the inefficiency of yard movements. In this research, the incoming container will

be placed temporary on the handshake area and

will be moved later to the landside or seaside area by the remarshaling jobs. There is no special

area dedicated to containers from certain vessels

(such as in consignment strategies), but containers from the same vessel will always be

placed close to each other. Containers from other

vessels will tend to be placed in different

locations at the yard. This strategy will minimized delay that is caused by rehandling

job.

3.2 Real-time scheduling of ASCs

The algorithm of real-rime ASCs

scheduling used in this research is developed from Algorithm for Opportunistic Remarshaling

(AOR) by Choe et al. [11]. Firstly, the ASC

scheduling method sorts out all the landside and

seaside jobs within a look-ahead horizon and finds out whether preparatory jobs (rehandling,

repositioning) are needed. Then the method

performs an optimization search to specify the ASC to perform each of the preparatory jobs, the

order of jobs, and the priority in case of ASC

interference. AOR adds some remarshaling jobs

for the upcoming vessel to be done at any idle time in-between the ordinary jobs. Due to

accomodating the uncertainty of the

environtment, the outdated schedule is refreshed by iterate this process at short regular intervals.

We follows the iterative recheduling method by Choe et al. [16] in which the process

of developing a new schedule is repeated in a

predetermined recheduling interval r as indicated in the timing diagram of Figure 3. In the

diagram, t_1 is the starting time of scheduling for

the jobs in the i-th iteration. The scheduling is made for the next look-ahead horizon h, within s

permittable computational time. Therefore, the

schedule period is between t_1+s until t_1+s+h.

The rescheduling process will be repeated every interval r. The length of rescheduling interval

and look-ahead horizon should be determined

carefully considering various tradeoffs [16]. As h gets longer, a better scheduling result might be

expected because a greater number of future jobs

are taken into account. However, longer h demands a larger amount of computational times

to generate a schedule, also it is not desirable

because the uncertainty at the operational level

regarding ET arrival time is very high, thus make the schedule obtained has poor quality. As in the

Choe et al. [11], we determine the values of r,h,s

are 300, \1800, and 60 seconds respectively.

3.2.1 Job Priority Determination

Before determining the priority, the first

step of this method is determining the jobs to be scheduled. The jobs can be divided into main

jobs and preparatory jobs. The main jobs include

seaside jobs which the expected AGV arrival times are within the look-ahead horizon



e-ISSN 2477-6025

DOI:10.21776/ub.jemis.2018.006.02.2

79

Fig 3. Timing Diagram of Iterative Rescheduling [11] and landside jobs which the ETs have already

arrived at the landside H/P. The arrival of ET is

randomly generated to reflect the uncertainty in

real system. The preparatory jobs included in the scheduling are the rehandling and repositioning

jobs required to be done in order to access the

container of the main jobs. If there is still time left on the look-ahead horizon, some

remarshaling jobs for upcoming vessel will be

scheduled. The seaside jobs are given top priority among all types of jobs, because it is

critical of causing the loading delays and

affecting the port performance.

For sorting out the list of jobs, firstly we calculate the estimate finish time of each job. To

maintain the preparatory jobs are done right

before their main jobs, we pair them into a single consecutive job and sum up their processing

time. For each determined job, we calculate the

difference between expected arrival time of ET/AGV/vessel and the estimate finish time of

the job. Then we sort jobs by giving priority to

jobs that have the smallest difference. The

overall scheduling process can be seen in the flowchart in Figure 4.

3.2.1 Determination of Stacking Locations This procedure is used to determine the

optimal location for container handled in the

yard. Several policies to determine the optimal

location is modified from Choe et al. [11], which

is also modification of previous work by Jang et

al. [17]. This stacking policy is required

whenever an ASC has to drop a container at

some location in the yard. The drop off location can be an empty ground in the block or at the top

off every stack which has not reached the

stacking limit. As for the stack which has already reached its allowed maximum, the stack has to

be filtered out from the list of candidate

positions. The stacking location should be determined not only for incoming containers but

also for rehandled or remarshaled containers. In

this research we limit the problem only for 20ft

container and ignore the container weight stacking rule.

We modified the scoring function of

Choe et al. [11] by adding function to calculate container similarity. The container similarity

function is used to direct container to locations

where other containers with similar destinations are located. By implementing this function, the

need for preparatory jobs will be minimized

because all containers in the same stacking

location have similar destination vessel. We eliminate the transshipment function because

transshipment process is not considered in this

study. The stacking rules for different types of containers in different situations can be seen in

Table 1. Each rules of the policy in the table has

a scoring function. Given a set of candidate



e-ISSN 2477-6025

DOI:10.21776/ub.jemis.2018.006.02.2

80

posisitions X, scoring function 𝑠𝑖 will determine the

Fig 4. Flowchart of Scheduling Process in CY

best position 𝑥∗from X. The score 𝑠𝑖(𝑥) of a

candidate posisition x is calculated by the weighted

sum function below. Where 𝐶𝑖𝑗(𝑥) is the evaluation

value of position x according to the rule of 𝑠𝑖. The

weight of 𝐶𝑖𝑗 is 𝑤𝑖𝑗 and 𝑘𝑖 is the number of criteria

employed by 𝑠𝑖.

𝒔𝒊(𝒙) = ∑ 𝒘𝒊𝒋𝑪𝒊𝒋(𝒙)

𝒌𝒊

𝒋=𝟏

( 1 )

Table 1. Clasification of Containers Stacking Rules

Condition Types of

Containers

Evaluation

Criteria

Incoming Import 𝑠1

Export 𝑠2

Rehandling Import 𝑠3

Export 𝑠4

Remarshaling Import 𝑠5

Start

1. Make a list of activities

2. Specify the location of containers

3. Calculate the process time and

specify the critical time

4. Sort the activities based on the

critical time

5. Input the activities in planning

horizon h

6. Is the

cumulative process

time < planning

horizon h?

7. Make improvement of the job

sequence to minimize total makespan

End

Data Random of the

Trucks Arrival

Data of the AGV

Arrival

Data of

Remarshaling

Yes

No



e-ISSN 2477-6025

DOI:10.21776/ub.jemis.2018.006.02.2

81

Export 𝑠6

Evaluation criteria used in the scoring function

could better be understood by examining the scoring functions more detail. Following is detail scoring

function of S2 used for incoming export containers:

𝒔𝟐 =𝒘𝟏𝑫𝟏(𝑿) +𝒘𝟐𝑫𝟎(𝑿) +𝒘𝟑𝑯(𝑿) +𝒘𝟒𝑻(𝑿) +𝒘𝟓𝑲(𝑿)

(2)

Where 𝐷1 is the distance of the current

position to candidate position X. 𝐷0(𝑋) is the distance of candidate position X from landside

H/P. 𝐻(𝑋) is the stack height of candidate

position. 𝑇(𝑋) is the likelihood of causing

rehandling if target container is put on candidate

position X. While 𝐾(𝑋) is the uniformity score

of the candidate position X in which represents

the degree of destination similarity with other containers surround it. For all candidate

positions X, the value of 𝑠2 is calculated then the

candidate position with smallest value is chosen

as the optimal position for the target container.

3.2.1 Searching for an Optimal Schedule After the job-queue for a look-ahead horizon

is fixed, the schedule then optimized by swapping the

sequence of the jobs to find a combination which gives minimum makespan. The objective function for

schedule optimization is the following weighted sum

that should be minimized:

𝒇 = 𝒘𝟏𝑫𝑨𝑮𝑽 + 𝒘𝟐𝑫𝑬𝑻 + 𝒘𝟑𝑴 +𝒘𝟒𝑹(3)

where 𝐷𝐴𝐺𝑉 is the average delay of AGVs, 𝐷𝐸𝑇 is the average waiting time of ETs, M is the

makespan, and R is the rehandling jobs. The

minimation of 𝐷𝐴𝐺𝑉, 𝐷𝐸𝑇 , and R are obtained in

generating jobs phase, while the minimation of makespan is generated from second phase,

which is improvement of job queue by swapping

the sequence.

4. Experimental Results

The experiment involving a set of vessel arrival data, random ET arrival data, and generated

AGV arrival data. The vessel data includes

information of date of vessel arrival and the number

of containers being loaded or unloaded from the ship.

The AGV arrival data generated within the window of vessel arrival and departure, the number of trips is

adjusted by the number of containers being carried,

both for loading and unloading containers. The available information is AGV arrival time to the yard,

type of operation, container code that being handled,

and vessel code that being served. Different from AGV, the random arrival data for ET to deliver the

container is spread between 5 consecutive days prior

the vessel departure. While the arrival of ET for

retrieving the import containers is generated randomly after vessel arrival. The information of ET

arrival also includes ET arrival time, type of

operation, container code that being handled, and vessel code of container destination. Those data are

compiled into Input Job data as the input of the

scheduling process. Table 2 represents several input

jobs for this experiment.

We set the scheduling parameter of r, h, s is

300, 1800, and 60 seconds respectively. The

algorithm first generated the queue job by identifying the arrival of ET and AGV within the look-ahead

horizon to be scheduled as the main jobs. When ET

or AGV brings in container to the yard, then the algorithm will find the optimum posisition on the

yard by implementing the scoring function. On the

other hand, if ET or AGV will takes out container

from the yard, the algorithm will find the location of targeted container and identify whether some

preparatory jobs are needed.

Some preparatory jobs (i.e rehandling and repositioning jobs) are needed if the targeted

container is located not on the top of the any stack. In

order to access the container, some containers above the targeted location have to be moved to other

location first. The destination location of rehandling

and repositioning container also being optimized

using rule number 3 and 4 on the Table 1. The rules are implemented to prevent any other movement of

that container in the future. The original targeted

container can be moved to ET or AGV by an ASC after all its preparatory jobs finished. The processing

time of the jobs become longer because of the

additional preparatory jobs. Table 3 shows the job that needed some preparatory jobs. An ET need to carry

out a container “MRLU2363500” from the location

of slot 1,3,2 at the yard. Unfortunately, there are three

containers above that have to be moved first before accessing the targeted container. The schedule will

determine the best location for the rehandling

containers, which are on the neighboring slot. Finally,



e-ISSN 2477-6025

DOI:10.21776/ub.jemis.2018.006.02.2

82

the targeted container can be moved to the landside area after all

Table 2. Input Jobs

Arrival Date Vehicle Activity Vessel ID Container ID

1/1/2018 0:02 ET 1 RELI006 MRTU2149246

1/1/2018 1:27 ET 1 RELI006 MRTU2045641

1/1/2018 1:30 ET 1 RELI006 TGHU0572960

1/1/2018 1:46 ET 1 RELI006 MRTU2042643

1/1/2018 3:05 AGV 1 RERO006 MRLU2387707

1/1/2018 3:10 AGV 1 RERO006 MRLU2363860

1/1/2018 3:15 AGV 1 RERO006 TEXU2582173

1/1/2018 3:20 AGV 1 RERO006 MRLU2341265

1/1/2018 3:25 AGV 1 RERO006 MRLU2380451

1/1/2018 3:25 ET 1 RELI006 TRLU9164682

1/1/2018 3:30 AGV 1 RERO006 MRTU9601872

1/1/2018 3:35 AGV 1 RERO006 MRTU9614904

1/1/2018 3:40 AGV 1 RERO006 MRLU4200287

1/1/2018 3:45 AGV 1 RERO006 MRTU2124891

1/1/2018 3:49 ET 0 RERO006 MRTU2059796

1/1/2018 3:50 AGV 1 RERO006 MRTU2059796

1/1/2018 3:55 AGV 1 RERO006 MRTU2155022

1/1/2018 4:00 AGV 1 RERO006 MRTU2054752

1/1/2018 4:05 AGV 1 RERO006 MRLU2363500

Table 3. Iterative Scheduling Result

NoContainer

NameOrigin Destination

Processing

Time

Finishing

Time

Arrival Time

at HPDeviation Information

1 MRLU2335880 Slot 1,3,5 Slot 1,4,1 31.1 Rehandling MRLU2363500

2 MRTU2022776 Slot 1,3,4 Slot 1,4,2 31.1 Rehandling MRLU2363500


4 MRLU2363500 Slot 1,3,2 Landside 22.1 1/1/2018 4:26 1/1/2018 4:25 -0.000856 Target Rehandling

5 MRLU2320771 Seaside Slot 18,1,1 55.9 1/1/2018 4:25 1/1/2018 4:25 -0.000637 In AGV

6 MRTU2072165 Seaside Slot 18,1,2 50.7 1/1/2018 4:25 1/1/2018 4:30 0.0028935 In AGV


8 MRTU2096927 Landside Slot 22,1,1 55.9 1/1/2018 4:25 1/1/2018 4:35 0.0065625 In ET





13 MRLU2335880 Slot 1,4,1 Landside 27.3 1/1/2018 4:25 1/1/2018 4:53 0.0189583 Target Rehandling

14 MRTU9608157 Slot 18,1,5 Slot 1,4,2 43.9 1/1/2018 4:25 1/1/2018 3:03 -0.057442 Remarshaling




18 MRLU2320771 Slot 18,1,1 Slot 1,5,1 38.7 1/1/2018 4:25 1/1/2018 3:03 -0.057384 Remarshaling


20 MRTU2096927 Slot 22,1,1 Slot 39,6,4 33.5 1/1/2018 4:25 1/2/2018 23:19 1.7871181 Remarshaling



e-ISSN 2477-6025

DOI:10.21776/ub.jemis.2018.006.02.2

83

Fig 5. Initial Layout before Rehandling

Fig 6. Layout after Rehandling

the rehandling jobs handled. The movement of

containers can be seen on figure 5 (initial layout) and

figure 6 (after rehandling layout).

After all main jobs in the look-ahead horizon

have been scheduled, the algorithm investigates for

the remaining time available. Some remarshaling jobs are added one by one to the queue until the

scheduling time is up. To find the best candidate

location for the remarshaling container, the algorithm will implement rule number 5 and 6 on Table 1

depend on the container type. The remarshaling job

for the export container will be handled by seaside

ASC, while remarshaling for import container handled by landside ASC. The remarshaling jobs will

only be done if there is container on the transfer area

and the scheduling time is still available. If the ASCs

are busy with the main jobs, then the remarshaling jobs will not be handled. In the given scenario, the

quantity of the job is not many, makes it sufficient to

handle the remarshaling jobs in the idle time of each scheduling window. The remarshaling jobs is marked

with “Remarshaling” title in information column of

Table 3. The position destination slot will always change during the iterative rescheduling because of

any movements from previous actions. The exact

location can only be obtained several minutes prior to

the processing time.

5. Conclusion



e-ISSN 2477-6025

DOI:10.21776/ub.jemis.2018.006.02.2

84

Yard management is important to improve the

operational performance of a port. Some important

factors on yard management are yard template generation and ASC scheduling. Generating yard

template has to be robust to accommodate the

uncertainties of real operational situations, such as the arrival of external trucks or containers. To

accommodate the uncertainties, some ports

implement the consignment strategy in which shippers book some location for their upcoming

containers. Practically, this strategy can improve the

accessibility of the container because container with

similar detination vessel will be placed nearch each other, but in some occasions this strategy will lead to

inefficient usability of yard area. This low utilization

of the yard area is caused by the uncertainty of container arrivals. In this research, the algorithm for

improving container yard utilization has been

developed and the conducted numerical experiment

showed some results, which are:

a. In the algorithm, the utilization is maintained

by placing the incoming container at the

transfer area first and rearranging it again near the similar vessel destination containers

through remarshalling process prior to the

departure date. b. The scheduling process in this research are

conducted in two steps, the first step is

determining the jobs queue and next step is

improving the sequence of the job to minimize the makespan.

c. The improvement process is done manually by

swapping the sequence of the queuing jobs. d. Some scoring rules that have been

implemented to determine the optimal location

for the containers represent a good performance by placing the similar containers

on the neighbouring area.

Future research is needed to accommodate

several characteristics of containers, including size and weight. The development of the improvement

method is also needed, for example by employing

heuristics algorithm rather than arbitrary swapping.

6. Acknowledgment We are thankful to the Direktorat Riset dan

Pengabdian Masyarakat - Direktorat Jendral Penguatan Riset dan Pengembangan - Kementrian Riset, Teknologi dan Pendidikan Tinggi - Indonesia for providing the funding for this research. We also thank all the staff in Lembaga Penelitian dan Pengabdian Masyarakat - Universitas Internasional Semen Indonesia for

facilitating this research.

References [1] The World Bank..Global Ranking 2018.

Available from:

https://lpi.worldbank.org/international/global [Accessed 17 September 2018].

[2] BAPPENAS. Laporan Implementasi Konsep

Tol Laut 2015-2019, Jakarta: Kementerian Perencanaan dan Pembangunan Nasional; 2015.

[3] Pitoko, R.A..Penasaran "Dwelling Time" di Atas

3 Hari, Menhub akan Datangi Tanjung Priok. Available from: https://ekonomi.kompas.com/read/2018/04/03/1

45600326/penasaran-dwelling-time-di-atas-3-

hari-menhub-akan-datangi-tanjung-priok. [Accessed 12 Nopember 2018].

[4] Nurminarsih, S., Rusdiansyah, A. and Putri,

M.M.. Space-sharing Strategy for Building

Dynamic Container Yard Storage Considering Uncertainty on Number of Incoming Containers.

Jurnal Teknik Industri. 2018;19(2): pp.67-74.

[5] Zhen, L.. Container yard template planning under uncertain maritime market.

Transportation Research Part E. 2014; 69: pp.

199-217. [6] Lee, L.H., Chew, E.P., Tan, K.C., Han, Y. An

optimization model for storage yard

management in transshipment hubs. OR

Spectrum. 2006; 28: pp. 539–561. [7] Jiang, X., Lee, L.H., Chew, E.P., Han, Y., Tan,

K.C. A container yard storage strategy for

improving land utilization and operation efficiency in a transshipment hub port. European

Journal of Operation Research. 2012; 221: pp.

64-73. [8] Alcalde, E.M., Kim, K.H., Marchan, S.S..

Optimal space for storage yard considering yard

inventory forecasts and terminal performance.

Transportation Research Pert E. 2015; 82: pp. 101-128.

[9] Han, X.L., Lu, Z.Q., Xi, L.F. A proactive

approach for simultaneous berth and quay crane scheduling problem with stochastic arrival and

handling time. European Journal of Operation

Research. 2010; 207: pp. 1327-1340.

[10] Zhen,L., Lee, L.H., Chew, E.P. A decision model for berth allocation under uncertainty. European

Journal of Operation Research. 2011; 212: pp.

54-68. [11] Choe, R., Kim, T.S, Kim, T.,Ryu , K.R. Crane

Scheduling for Opportunistic Remarshaling of

Containers in an Automated Stacking Yard.



e-ISSN 2477-6025

DOI:10.21776/ub.jemis.2018.006.02.2

85

Flexible Service and Manufacturing Journal.

2015; 27: pp 331-349.

[12] Park T, Kim J, Ryu KR. Iterative replanning using genetic algorithms for remarshaling in a

container terminal. InProceedings of the 10th

IASTED International Conference. 2010; Vol. 674, No. 120, p. 22.

[13] Park K, Park T, Ryu KR. Planning For Selective

Remarshaling in an Automated Container Terminal using Coevolutionary Algorithms.

International Journal of Industrial Engineering.

2013 Jan 1;20.

[14] Park T, Choe R, Ok SM, Ry KR. Real-time scheduling for twin ASCs in an automated

container. OR Spectr. 2010; 32: pp593-615.

[15] Carlo, H.J., Vis, I.F.A., Roodbergen, K.J. Storage yard operations in container terminals:

literature review, trends, and research directions.

European Journal of Operation Research 2014: 235: pp 412-430.

[16] Choe R, Yuan H, Yang Y, Ryu KR. Real-time

scheduling of twin stacking cranes in an

automated container terminal using a genetic algorithm. InProceedings of the 27th Annual

ACM Symposium on Applied Computing 2012

Mar 26. pp. 238-243. ACM. [17] Jang H, Choe R, Ryu KR. Deriving a robust

policy for container stacking using a noise-

tolerant genetic algorithm. InProceedings of the

2012 ACM Research in Applied Computation Symposium 2012 Oct 23. pp. 31-36. ACM.

DYNAMIC YARD TEMPLATE DESIGN TO IMPROVE UTILIZATION …

Documents

Transcript of DYNAMIC YARD TEMPLATE DESIGN TO IMPROVE UTILIZATION …